Bob Pette, NVIDIA & Neil MacDonald, HPE | HPE Discover 2024
summary
Announcement details a co-developed, turnkey stack called NVIDIA AI Computing by HPE that packages HPE ProLiant Gen12 servers, GreenLake storage, integrated NeMo retriever and NVIDIA runtimes so enterprises can fine-tune models, run RAG pipelines and deploy inference in under 30 seconds. It uses containerized NIMs (inference microservices) validated nightly against replicated systems, supports small–extra-large footprints, targets on‑prem private cloud with liquid cooling, and maps to a one‑year GPU cadence (Hopper → Blackwell → Rubin).
AI-written summary · about
chapters (12)
- 0:00HP and Nvidia partnership
- 1:11AI technology deployment
- 2:43Enterprise AI systems
- 5:14Private Cloud for AI
- 6:59Model and data stack
- 8:58Transforming business data
- 12:45Intention-based computing
- 16:01Hardware roadmap
- 17:47Cuda and future innovations
- 20:21Efficiency and performance
- 23:03Liquid cooling technology
- 25:52Go-to-market strategy
